Am 08.01.2024 um 18:32:25 Uhr schrieb The Natural Philosopher:

> In every case where I might have wanted to use it, the overhead of
> transmitting graphics over the network killed it. Far easier to send
> the underlying data over the network and have a fat graphics client
> like a web browser at the other end...

Although that works fine motif-based applications, but for Qt or GTK,
it is still nasty.

At least, ssh provides -C to compress all the stuff that massively
reduced the load.

The Natural Philosopher <tnp@invalid.invalid> wrote:
> On 08/01/2024 14:37, Marco Moock wrote:
>> Am 08.01.2024 um 12:42:03 Uhr schrieb The Natural Philosopher:
>>
>>> Looking at it fairly cursorily, it looks like a Wayland core with an
>>> x windows shim *if needed* is actually a better way to go long term,
>>> as X-windows (like PostScript) is full of code that 99% of people
>>> never use.
>>
>> What is the problem with that?
>>
> it introduces a huge CPU, memory, programming and support overhead for
> something that is (almost) never used.

A lot of the features that are less used are actually performance
optimisations. Some might find them irrelevent because they only
use extremely powerful PCs these days, but many protocol features
and extensions were intended to reduce performance overhead, not
increase it. Plus it's all designed to be flexibly configured (and
"modular", like the Linux kernel itself), so you can avoid a lot of
the allegedly useless code anyway.

Performance with Wayland, so far as I could see, only seemed to be
considered in terms of 3D graphics such as for commercial video
games. Similar performance for general desktop tasks doesn't seem
to be a consideration. Wayland's about writing a graphics system
that Red Hat and friends find it easier to work on. Xorg's only
gotten fatter and slower with their work over the last decade or so
anyway, so I'm glad t have the _continuing_ development work on
Xorg done without them anyway.

I talk about Wayland in the past tense because I saw that it had
nothing new to offer me for my usage, while failing to bother with
replicating X's fabulous support for running remote windows over
the network. Xorg still has development going on, and older
implementations such as XFree86 and the TinyX sub-project can still
be used after development has been abandoned for over a decade. The
only trouble I might have is needing to run Wayland displaying on X
if Wayland-only software starts coming out. But actually most
software I run is X-only anyway and unlikely to ever support
Wayland because nobody's touched its code in decades either.

>> I like X11, it does what it should and can be customized.
>>
> It's been buggered till it works OK, sure...
>
>> From Wayland, I only hear about problems with graphic drivers (mostly
>> Nvidia's crappy blob).
>
> Ok, that's the sort of thing I was interested in...

Nvidia's drivers had a bad rep. on X as well often enough. Nvidia
GPUs seem to me like a bad option for Linux users in general.

Drivers are probably where Wayland will soon have the advantage.
Maybe they'll be slow and buggy, but companies that used to
contribute to GPU drivers for X (including open-source code
contributors like Intel) will now be focusing on Wayland because
that's where Red Hat's pointing them.

That said, the movement of GPU driver code from the display server
to the Linux kernel with DRM means that theoretically it should be
easy for the (_still_ _existant_) Xorg developers to support the
same drivers in X. In practice it might depend on whether anyone
upstream listens if they complain about a bug that only affects X
and not Wayland.

But at a minimum the Linux framebuffer is always a fall-back option
which works plenty fast enough for basic desktop graphics tasks and
should continue to work in X for the forseeable future. Indeed until
recently that's what all the Raspberry Pi boards relied on for
their graphics since an open-source GPU driver was very slow to be
realised for them.

--
__ __
#_ < |\| |< _# | Note: I won't see posts made from Google Groups |

Am 08.01.2024 um 16:44:12 Uhr schrieb Blue-Maned_Hawk:

> Wayland has already obsolesced X11 and will eventually completely
> eliminate it everywhere. There is no stopping this.

Why do you think?
Many systems still run X11 and the nvidia blob doesn't support it (yet).

Am 08.01.2024 um 17:46:58 Uhr schrieb The Natural Philosopher:

> - It is not as full featured as X11, but there is no compatibility
> problem with running X11 on top of Wayland, as is done in Mac OS-X.

Isn't that a crappy solution?
All of the stuff that uses X11 need to adapt that.
Does that also support simple window manager like twm or mwm?

Am 08.01.2024 um 12:32:29 Uhr schrieb David W. Hodgins:

> On Mon, 08 Jan 2024 07:42:03 -0500, The Natural Philosopher
> <tnp@invalid.invalid> wrote:
> > Looking at it fairly cursorily, it looks like a Wayland core with
> > an x windows shim *if needed* is actually a better way to go long
> > term, as X-windows (like PostScript) is full of code that 99% of
> > people never use.
> >
> > So on the surface it looks like Wayland would be faster and use a
> > lot less memory.
> >
> > I am sure however that there are hidden gotchas.
> >
> > So: interested to hear from everyone.
>
> While Xorg is a security nightmare that is no longer being
> maintained,

From where do you have that information?

> it does still work and the security aspect can be
> mitigated with proper network security.

By default, the XServer isn't reachable from the network interface
because it doesn't listen there.
I know this was different in the past.

X11 forwarding can be securely done with SSH.

> I use both X11Forwarding over ssh and synergy, so am sticking with
> Xorg for now.

I also use X11 forwarding and I also use mwm (I don't like GNOME or
KDE).
Is something similar available for Wayland?

Am 08.01.2024 um 17:55:37 Uhr schrieb The Natural Philosopher:

> On 08/01/2024 17:32, David W. Hodgins wrote:
> > On Mon, 08 Jan 2024 07:42:03 -0500, The Natural Philosopher
> > <tnp@invalid.invalid> wrote:
> >> Looking at it fairly cursorily, it looks like a Wayland core with
> >> an x windows shim *if needed* is actually a better way to go long
> >> term, as X-windows (like PostScript) is full of code that 99% of
> >> people never use.
> >>
> >> So on the surface  it looks like Wayland would be faster and use a
> >> lot less memory.
> >>
> >> I am sure however that there are hidden gotchas.
> >>
> >> So: interested to hear from everyone.
> >
> > While Xorg is a security nightmare that is no longer being
> > maintained, it does still work and the security aspect can be
> > mitigated with proper network
> > security.
> >
> Is Xorg the bulk of the X11 windowing API?

X11 is a protocol/specification.
Certain implementations exist.

Proprietary ones on the UNIX system like AIX, Solaris and free ones.
XFree86 was a free port on the 386 CPU line.
Later that project stalled and was forked as X.org.

The Natural Philosopher <tnp@invalid.invalid> wrote:
> Is Xorg the bulk of the X11 windowing API?

Xlib is the API, and Xorg is the most actively developed
implementation of that since it forked from XFree86. Xorg's Xlib
talks to the X server via the X11 protocol, and there are more
alternative implementations of the X server for running on other
OSs such as Windows and Mac.

The one other open-source Xlib implementation that I know of which
is still seeing some slight development activity is in Nano-X:
http://www.microwindows.org

Some X11 software will compile+run OK against Nano-X's Xlib
equivalent, but not all.

There's an old distro called Nano Linux which used Nano-X. It's
pretty buggy, but it's more interesting to me than Wayland because
it's truely small and lightweight:
https://sourceforge.net/projects/nanolinux/

--
__ __
#_ < |\| |< _# | Note: I won't see posts made from Google Groups |

Computer Nerd Kev <not@telling.you.invalid> wrote:
> The Natural Philosopher <tnp@invalid.invalid> wrote:
>> Is Xorg the bulk of the X11 windowing API?
>
> Xlib is the API, and Xorg is the most actively developed
^^
I should have said "includes", the key point being that the library
and the server are separate parts of Xorg.

--
__ __
#_ < |\| |< _# | Note: I won't see posts made from Google Groups |

On 1/8/24 13:42, The Natural Philosopher wrote:
> Looking at it fairly cursorily, it looks like a Wayland core with an x
> windows shim *if needed* is actually a better way to go long term, as
> X-windows (like PostScript) is full of code that 99% of people never use.

I fear that code is there for a reason, and Wayland throws the baby out
with the bathwater.

Did you know that Wayland clients aren't allowed to know or set their
window coordinates on the screen?

>
> So on the surface  it looks like Wayland would be faster and use a lot
> less memory.
>
> I am sure however that there are hidden gotchas.

When you remove a bunch of unused code, there usually are.

https://www.joelonsoftware.com/2000/04/06/things-you-should-never-do-part-i/

>
> So: interested to hear from everyone.
>

On 1/8/24 17:46, Blue-Maned_Hawk wrote:
> Marco Moock wrote:
>
>> From Wayland, I only hear about problems with graphic drivers (mostly
>> Nvidia's crappy blob).
>
> That is the fault of nvidia, not the fault of Wayland.
>

Actually everything in open source is the fault of open source people.
It's not like Nvidia has any obligation to support non-Windows
platforms. We wish they would, but they don't have to.

On 1/8/24 17:44, Blue-Maned_Hawk wrote:
> Wayland has already obsolesced X11 and will eventually completely
> eliminate it everywhere. There is no stopping this.

2024 is the year of the wayland desktop.

On 1/8/24 18:46, The Natural Philosopher wrote:
> Again, I am not in possession of hard facts, but what it *seems* to be
> like is:
>
> - Wayland can replace X11 between  a graphics API (Gnome etc) with
> potentially a lower footprint and CPU usage for a faster response for
> most desktop applications..

Why do you think so? Most X11 systems today don't transmit graphics
through X. They render directly to GPU buffers using DRM (direct
rendering manager, not the other thing), just like in Wayland.

Marco Moock <mm+usenet-es@dorfdsl.de> wrote:
> Am 08.01.2024 um 17:46:58 Uhr schrieb The Natural Philosopher:
>
>> - It is not as full featured as X11, but there is no compatibility
>> problem with running X11 on top of Wayland, as is done in Mac OS-X.
>
> Isn't that a crappy solution?
> All of the stuff that uses X11 need to adapt that.
> Does that also support simple window manager like twm or mwm?

My understanding of Wayland compositors is that they need to
include more logic to do some functions equivalent to some roles
of the X server itself in the X Window system. Hence window
managers and compositors tend to be developed as separate projects
because there's too much difference between their functionality.
Apparantly some clones of popular window managers have been made
as Wayland compositors.

As they don't use graphics toolkits that support Wayland, TWM and
MWM are examples of software that would only work in Wayland via
XWayland even if they weren't Window Managers. For window managers,
well you need a compositor to run Wayland+XWayland, and the
compositor takes on the role of a window manager, so the idea of
using an X window manager in Wayland via XWayland is a logical
contradiction. Anyway if they did at least talk Wayland via a new
X11/Wayland graphics toolkit, that still wouldn't give them any of
the extra functionality required from a Wayland compositor.

--
__ __
#_ < |\| |< _# | Note: I won't see posts made from Google Groups |

On 1/8/24 23:06, David W. Hodgins wrote:
> On Mon, 08 Jan 2024 12:55:37 -0500, The Natural Philosopher
> <tnp@invalid.invalid> wrote:
>> Is Xorg the bulk of the X11 windowing API?
>
> Both xorg and wayland run under X11.
>
> It's xorg that is being replaced by wayland.
>
> Another example of a feature I find missing in wayland is the redshift.
> As xorg allows any application to access or modify other applications
> output,
> while wayland does not, redshift only works on xorg.
>
> It's the design of xorg that makes it inherently insecure, but easier to
> use
> when intentionally using things like reshift and synergy.
>
> Regards, Dave Hodgins

Indeed, I have never wished that the programs on my desktop could
interact with each other less, and if I do want to run an untrusted
program in a sandbox, it has to be a full sandbox - just sandboxing the
windowing system is far from sufficient.

On 08/01/2024 20:55, Marco Moock wrote:
> Am 08.01.2024 um 17:46:58 Uhr schrieb The Natural Philosopher:
>
>> - It is not as full featured as X11, but there is no compatibility
>> problem with running X11 on top of Wayland, as is done in Mac OS-X.
>
> Isn't that a crappy solution?
> All of the stuff that uses X11 need to adapt that.

The whole point of that is that they wont

> Does that also support simple window manager like twm or mwm?
>
If you are in X11 mode, I guess so.

--
New Socialism consists essentially in being seen to have your heart in
the right place whilst your head is in the clouds and your hand is in
someone else's pocket.

On 08/01/2024 21:43, immibis wrote:
> On 1/8/24 18:46, The Natural Philosopher wrote:
>> Again, I am not in possession of hard facts, but what it *seems* to be
>> like is:
>>
>> - Wayland can replace X11 between  a graphics API (Gnome etc) with
>> potentially a lower footprint and CPU usage for a faster response for
>> most desktop applications..
>
> Why do you think so? Most X11 systems today don't transmit graphics
> through X. They render directly to GPU buffers using DRM (direct
> rendering manager, not the other thing), just like in Wayland.

If so that validates my possible point that large chunks of X112 are
redundant...

Oh well I am sure that wayland will appear on one of my machines in due
course..

--
"Strange as it seems, no amount of learning can cure stupidity, and
higher education positively fortifies it."

- Stephen Vizinczey

On 1/9/24 00:59, The Natural Philosopher wrote:
> On 08/01/2024 21:43, immibis wrote:
>> On 1/8/24 18:46, The Natural Philosopher wrote:
>>> Again, I am not in possession of hard facts, but what it *seems* to
>>> be like is:
>>>
>>> - Wayland can replace X11 between  a graphics API (Gnome etc) with
>>> potentially a lower footprint and CPU usage for a faster response for
>>> most desktop applications..
>>
>> Why do you think so? Most X11 systems today don't transmit graphics
>> through X. They render directly to GPU buffers using DRM (direct
>> rendering manager, not the other thing), just like in Wayland.
>
> If so that validates my possible point that large chunks of X112 are
> redundant...
>

They are. Wayland removes chunks that are redundant and chunks that are
not redundant.

It's also arguably desirable that X servers should support networked
drawing. Why should a display system require every client application to
possess a full GPU driver? Why can't I just send the server the pixels I
want to draw?

The Natural Philosopher <tnp@invalid.invalid> wrote:
> On 08/01/2024 16:24, Lew Pitcher wrote:
>> For what it's worth, I'm not going to get into a p*ssing match with you
>> about why X11 is better than Wayland. I don't care.
>>
>> If you want to expound on why Wayland is better than X11, go right ahead.
>> But, if you are truely interested in the opinion of others, how about
>> listening, rather than defending?
>>
> I am neutral. No dog in the fight.
>
> So far you have said 'remote desktop' and that's it.
> I tried a remote x desktop once. it was crap.

If by 'desktop' you mean a full environment like gnome, kde, or others,
then that is not likely to fully show the advantage of running remote X
programs.

Where X's network transparency becomes very useful is not in trying to
be a thin client to a remote fat desktop, but instead it is useful when
running your full desktop locally, then using X remote to run specific
X programs on specific remote machines, all while their windows appear
locally on your local desktop system.

I.e., to use X to run programs than need GUI's remotely in much the
same way that ssh allows you to run CLI programs on remote systems, yet
see their output appear locally.

RDP/VNC/etc. do not provide the same setup because they require the
remote system to at a minimum be running an X session with at least the
framebuffer console on a video card on the remote machine. [1] With
just running individual X programs remotely, the remote can be
headless, even to the point of no video card, and so long as all the X
library's are available, the remote program runs just as if it had a
local X server to use.

[1] Well, one 'could' run xnest on the remote to have a virtual
framebuffer on the remote for RDP/VNC to then forward.

On Mon, 08 Jan 2024 17:39:10 +0000, The Natural Philosopher wrote:

> On 08/01/2024 16:24, Lew Pitcher wrote:
>> For what it's worth, I'm not going to get into a p*ssing match with you
>> about why X11 is better than Wayland. I don't care.
>>
>> If you want to expound on why Wayland is better than X11, go right ahead.
>> But, if you are truely interested in the opinion of others, how about
>> listening, rather than defending?
>>
> I am neutral. No dog in the fight.
>
> So far you have said 'remote desktop' and that's it.

No. I did not say "remote desktop". I said "remote clients".

Here's an example: on my lan, I have three systems: "workstation", "webserver",
and "phoneserver". From a terminal on "workstation", I can
ssh -Y webserver "firefox http://localhost/"" rel="nofollow" target="_blank">http://localhost/"
which executes firefox on the "webserver", giving it the url http://localhost/

WHAT I SEE is a firefox window open up on my "workstation", browsing the
"webserver"'s localhost webpage.

This, you cannot do with Wayland.

> I tried a remote x desktop once. it was crap.

"Remote X desktop"? Do you mean RDP? This is /not/ RDP.
If you /want/ an RDP substitute, X11 can do that. Your "desktop manager"
is just another X11 client, and you can run it remotely just as easily
as the firefox example I give above. I even wrote a blog article about
it: http://justlinux.ca/node/82

--
Lew Pitcher
"In Skills We Trust"

On 2024-01-08, immibis <news@immibis.com> wrote:
> On 1/8/24 13:42, The Natural Philosopher wrote:
>> Looking at it fairly cursorily, it looks like a Wayland core with an x
>> windows shim *if needed* is actually a better way to go long term, as
>> X-windows (like PostScript) is full of code that 99% of people never use.
>
> I fear that code is there for a reason, and Wayland throws the baby out
> with the bathwater.
>
> Did you know that Wayland clients aren't allowed to know or set their
> window coordinates on the screen?

Are the rumors true that Wayland is hard-limited to the Doze-like
focus-on-top-window-only restriction?

Having used X since the 1980s, I strongly prefer
focus-follows-mouse and not click-to-both-raise-and-focus.

One of the _BIG_ usability advantages of X vs. Doze is X allows
keyboard focus on a non-top-most window. That makes it possible
to type into a window that is mostly obscured. It's a royal pain
to not have that ability on the Doze laptop I have to use for
work.

Thanks.

--
Robert Riches
spamtrap42@jacob21819.net
(Yes, that is one of my email addresses.)

On 1/9/24 06:04, Robert Riches wrote:
> On 2024-01-08, immibis <news@immibis.com> wrote:
>> Did you know that Wayland clients aren't allowed to know or set their
>> window coordinates on the screen?
>
> Are the rumors true that Wayland is hard-limited to the Doze-like
> focus-on-top-window-only restriction?

That sounds like compositor-defined behaviour.

Am 08.01.2024 um 22:39:55 Uhr schrieb immibis:

> On 1/8/24 17:46, Blue-Maned_Hawk wrote:
> > Marco Moock wrote:
> >
> >> From Wayland, I only hear about problems with graphic drivers
> >> (mostly Nvidia's crappy blob).
> >
> > That is the fault of nvidia, not the fault of Wayland.
> >
>
> Actually everything in open source is the fault of open source
> people. It's not like Nvidia has any obligation to support
> non-Windows platforms. We wish they would, but they don't have to.

The good thing with FOSS is that anybody else who wants can create a
fork and make it compatible.

Am 08.01.2024 um 23:54:38 Uhr schrieb The Natural Philosopher:

> On 08/01/2024 20:55, Marco Moock wrote:
> > Am 08.01.2024 um 17:46:58 Uhr schrieb The Natural Philosopher:
> >
> >> - It is not as full featured as X11, but there is no compatibility
> >> problem with running X11 on top of Wayland, as is done in Mac
> >> OS-X.
> >
> > Isn't that a crappy solution?
> > All of the stuff that uses X11 need to adapt that.
>
> The whole point of that is that they wont
>
> > Does that also support simple window manager like twm or mwm?
> >
> If you are in X11 mode, I guess so.

But what is the benefit of using Wayland here?
It is just another application that needs to run.

Am 08.01.2024 um 17:17:51 Uhr schrieb David W. Hodgins:

> On Mon, 08 Jan 2024 15:57:32 -0500, Marco Moock

> > Is something similar available for Wayland?
>
> I've never used the Motif window manager.

I simply need a lightweight window manager that offers me windows that
can be moved, minimized etc.

On 09/01/2024 00:44, immibis wrote:
>Why can't I just send the server the pixels I
> want to draw?
Well you can. Its called an 'image'.

But of course you must be young not to remember why we didn't send
pixels and developed 'fat' clients: Network bandwidth.

IBMs original PC was possibly conceived in exactly that rôle... as
something smart to connect to an IBM mainframe.

--
WOKE is an acronym... Without Originality, Knowledge or Education.